• 한국멀티미디어학회논문지
• /
• 제22권4호
• /
• pp.432-442
• /
• 2019

Most current human action recognition methods based on deep learning methods. It is required, however, a very high computational cost. In this paper, we propose an action change detection method to reduce repetitive human action recognition tasks. In reality, simple actions are often repeated and it is time consuming process to apply high cost action recognition methods on repeated actions. The proposed method decides whether action has changed. The action recognition is executed only when it has detected action change. The action change detection process is as follows. First, extract the number of non-zero pixel from motion history image and generate one-dimensional time-series data. Second, detecting action change by comparison of difference between current time trend and local extremum of time-series data and threshold. Experiments on the proposed method achieved 89% balanced accuracy on action change data and 61% reduced action recognition repetition.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송∙미디어공학회 2018년도 하계학술대회
• /
• pp.84-87
• /
• 2018

Handwritten letter recognition is important for numerous real-world applications and many topics like human-machine interaction, education, entertainment, and more. This paper describes the implementation of a real-time handwritten letters recognition system on a common embedded computer. Recognition is performed using a customized convolutional neural network, which was designed to work with low computational resources such as the Raspberry Pi platform. The experimental results show that the proposed real-time system achieves an outstanding performance in the accuracy rate and the response time for recognition of twenty-six handwritten letters.

• 한국기계가공학회지
• /
• 제19권8호
• /
• pp.8-14
• /
• 2020

In the air current pulverizing type grinding method, the blade wings fitted inside a casing are rotated at a high speed to generate a cornering air current, which facilitates the collision of materials with one another, leading to the pulverizing phenomenon. In contrast to mechanical grinding, grit pulverizing leads to fine grinding and less acid waste and degeneration of the material. Moreover, this approach prevents the loss of nutritional value, while allowing the milling grain to have an excellent texture. However, the existing air current pulverizing type machines consist of prefabricated blades, which cannot be rotated at a speed higher than 5,000 rpm. Consequently, the grinding process becomes time consuming with a low productivity. To overcome these problems, in this study, the shape and structure of the air current pulverizing type wings were optimized to allow rapid grinding at more than 8,000 rpm. Moreover, the optimal design for the ripening parts for the air current pulverizing type device was determined by performing a computational fluid dynamics analysis based on airflow analyses to produce machinery that can grinding materials to the order of micrometers.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제7권8호
• /
• pp.1843-1859
• /
• 2013

The continuous increase of the spatial resolution of remote sensing images brings great challenge to image analysis and processing. Traditional prior knowledge-based region detection and target recognition algorithms for processing high resolution remote sensing images generally employ a global searching solution, which results in prohibitive computational complexity. In this paper, a more efficient region of interest (ROI) detection algorithm based on visual attention and threshold segmentation (VA-TS) is proposed, wherein a visual attention mechanism is used to eliminate image segmentation and feature detection to the entire image. The input image is subsampled to decrease the amount of data and the discrete moment transform (DMT) feature is extracted to provide a finer description of the edges. The feature maps are combined with weights according to the amount of the "strong points" and the "salient points". A threshold segmentation strategy is employed to obtain more accurate region of interest shape information with the very low computational complexity. Experimental statistics have shown that the proposed algorithm is computational efficient and provide more visually accurate detection results. The calculation time is only about 0.7% of the traditional Itti's model.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제5권2호
• /
• pp.408-427
• /
• 2011

Intensive computational complexity is an obstacle of enabling multiview video coding for real-time applications. In this paper, we present a fast macroblock (MB) mode selection algorithm for B frames which are based on the computational complexity analyses between the MB mode selection and reference frame selection. Three strategies are proposed to reduce the coding complexity jointly. First, the temporal correlation of MB modes between current MB and its temporal corresponding MBs is utilized to reduce computational complexity in determining the optimal MB mode. Secondly, Lagrangian cost of SKIP mode is compared with that of Inter $16{\times}16$ modes to early terminate the mode selection process. Thirdly, reference frame correlation among different Inter modes is exploited to reduce the number of reference frames. Experimental results show that the proposed algorithm can promote the encoding speed by 3.71~7.22 times with 0.08dB PSNR degradation and 2.03% bitrate increase on average compared with the joint multiview video model.

• 한국정보통신학회논문지
• /
• 제10권12호
• /
• pp.2315-2320
• /
• 2006

낮은 대역폭을 가지는 네트웍을 통해 비디오를 전송하기 위해 개발된 H.263과 같은 낮은 비트율 비디오는 대역폭에 맞게 출력 데이터를 조절하여야 한다. 본 논문에서는 토큰 버킷 알고리즘을 사용하여 입력 트래픽을 관리하는 네트웍에 적합한 비트율 제어 알고리즘을 제안한다. 제안하는 알고리즘은 계산의 복잡도를 줄이기 위해 반복 연산을 수행하지 않는 최 적화 방법을 사용하였고, 전체 시퀀스의 평균 왜곡을 최소화하면서도 인접 프레임 사이의 왜곡 차이를 최소화하도록 알고리즘을 설계하였다. 비디오 품질의 변동을 감소시키기 위하여 제안된 방법은 슬라이딩 윈도우 개념을 도입하였으며, 이 방법은 선분석(pre-analysis) 처리를 요구하지 않는 특징이 있다. 따라서, 제안된 알고리즘은 비디오를 압축할 때 추가적 인 지연을 필요로 하지 않게 되어 실시간 처리와 낮은 복잡도를 요구하는 비디오 압축기에 적용될 수 있다. 실험 결과에서 제안된 방법은 기존의 비트율 제어 알고리즘에 비해 영상의 품질이 우수한 것을 보여준다.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2003년도 추계학술대회
• /
• pp.127-132
• /
• 2003

A study of high-pressure n-heptane droplet vaporization is conducted with emphasis placed on equilibrium at vapor-liquid interface. General frame of previous rigorous model[1] is retained but tailored for flash equilibrium calculation of vapor-liquid interfacial thermodynamics. The model is based on complete time-dependent conservation equations with a full account of variable properties and vapor-liquid interfacial thermodynamics. The influences of high-pressure phenomena, including ambient gas solubility, thermodynamic non-ideality, and property variation on the droplet evaporation are investigated. The governing equations and associated moving interfacial boundary conditions are solved numerically using a implicit scheme with the preconditioning method and the dual time integration technique. And a parametric study of entire droplet vaporization history as a function of ambient pressure, temperature has been conducted. Some computational results are compared with Sato's experimental data for the validation of calculations. For low ambient temperatures, the droplet lifetime first increases with pressures, then decreases for high pressures. For higher ambient temperatures, the droplet lifetime increase with less amplitude than that of low ambient temperatures, which then decreases with more amplitude than that of low temperatures. The solubility of nitrogen can not be neglected in the high pressure and it becomes higher as the pressure goes up.

• 한국군사과학기술학회지
• /
• 제23권3호
• /
• pp.204-212
• /
• 2020

For general target tracking works by receiving a set of measurements from sensor. However, if the SNR(Signal to Noise Ratio) is low due to small RCS(Radar Cross Section), caused by remote small targets, the target's information can be lost during signal processing. TBD(Track Before Detect) is an algorithm that performs target tracking without threshold for detection. That is, all sensor data is sent to the tracking system, which prevents the loss of the target's information by thresholding the signal intensity. On the other hand, using all sensor data inevitably leads to computational problems that can severely limit the application. In this paper, we propose an iterative Joint Integrated Probabilistic Data Association as a practical target tracking technique suitable for a low SNR multi-target environment with real time operation capability, and verify its performance through simulation studies.

• ETRI Journal
• /
• 제38권1호
• /
• pp.100-111
• /
• 2016

This paper presents a low-complexity channel-adaptive reconfigurable $4{\times}4$ QR-decomposition and M-algorithm-based maximum likelihood detection (QRM-MLD) multiple-input and multiple-output (MIMO) detector. Two novel design approaches for low-power QRM-MLD hardware are proposed in this work. First, an approximate survivor metric (ASM) generation technique is presented to achieve considerable computational complexity reduction with minor BER degradation. A reconfigurable QRM-MLD MIMO detector (where the M-value represents the number of survival branches in a stage) for dynamically adapting to time-varying channels is also proposed in this work. The proposed reconfigurable QRM-MLD MIMO detector is implemented using a Samsung 65 nm CMOS process. The experimental results show that our ASM-based QRM-MLD MIMO detector shows a maximum throughput of 288 Mbps with a normalized power efficiency of 10.18 Mbps/mW in the case of $4{\times}4$ MIMO with 64-QAM. Under time-varying channel conditions, the proposed reconfigurable MIMO detector also achieves average power savings of up to 35% while maintaining a required BER performance.

• 한국CDE학회논문집
• /
• 제12권1호
• /
• pp.50-57
• /
• 2007

In recent years, many robotics researches have been focused on developing human-friendly robots, that is, humanoid biped robots. The researches of humanoid robots include various areas such as hardware development, control of biped locomotion, artificial intelligence, human interaction, etc. The present work concerns the hardware development of a mid-size humanoid robot, BONOBO, focusing on rapid development of outer body parts with integrated application if CAD/CAM/CAE/RP. Most parts are three-dimensionally designed using 3D CAD, and effectively connected with CAE analyses using both kinematic simulation and structural analysis. In order to reduce lead time and investment cost for parts developments, Rapid Prototyping (RP) and CAM are selectively utilized for manufacturing body parts. These master parts are then replicated using the vacuum casting process, from which we can obtain plastic parts repeatedly. Through this integrated approach, the first prototype of BONOBO can be successfully developed with relatively low time and investment costs.